About CityMapX
We turn city discussion scattered across social platforms and Google into a verifiable, shareable decision map. CityMapX isn't a food blog or a coupon site — it's a tool that helps you decide, giving you a clear pick while honestly flagging who it's not for and where the risks are.
How it works
From social noise to a decision you can walk out the door with, in four steps.
- 1
Gather community + Google
We pull together Google Reviews, PTT, and local blog discussion (Threads and Dcard coming soon), zeroing in on topics worth talking about.
- 2
Sort into four buckets
Split into factual rules, review spectrum, trust alerts, and recency signals — never blended into a single score.
- 3
Draw it as a smart map
Organized by "place × angle" into a browsable topic map, marking who it's good for and where the pitfalls are.
- 4
You verify, then decide
Every conclusion links to the original source and sample size so you can check it yourself — not an AI oracle you just take on faith.
What CityMapX shows you
Community review spectrum
Positive and negative reviews aren't blended into one star rating — a 5-axis spread shows where a place shines and where opinions diverge.
Taste · Service · Value · Ambiance · FreshnessMap comparison
Compare multiple places on the same topic side by side on a map — queues, timing, and price at a glance.
Place × AngleDecision shortcuts
Budget-friendly, local favorites, skip-the-rush picks... one tap filters to places that actually match.
Real filtering, not just tagsSource evidence
Every conclusion opens up to show platform breakdown, representative quotes, sample size, and confidence.
VerifiableAI doesn't just hand down a verdict
We turn raw reviews from community platforms and Google into judgments you can check: every conclusion shows its source platform, sample size, and confidence, with representative quotes you can open. Google's star rating is background reference only, never driving the ranking; when data is thin, we honestly label it "Curating" instead of making up numbers.
Data sources & methodology
How we pick, where the data comes from, and why it's verifiable.
- We cross-reference discussion from multiple sources — Google Reviews, PTT, local blogs — rather than relying on a single platform.
- We look at sentiment axes, sample size, and recency rather than blending them into one overall score; Google's star rating is background reference only and never drives ranking.
- When data is thin, we honestly label it "Curating / For reference only" instead of making up numbers.
- Every conclusion comes with its source platform, sample size, and confidence, plus representative quotes you can open and verify yourself.
- Currently integrated: Google Reviews, PTT, local blogs. Threads and Dcard are planned and will be added with in-app notice.
What we calibrate for
To make sure you see the negative reviews, we deliberately pull the "lowest rated" page when scraping, which over-represents negative reviews in the sample. We correct for this after the fact using the platform's own official rating distribution, pulling the ratio back toward the platform's overall distribution — but we only scale the intensity, never turn real negative reviews into positive ones.
What we can't fix (told to you straight)
- Platform rating inflation:Online ratings are inherently skewed toward the high end and the extremes (academics call it a J-shaped distribution); since our calibration baseline is also platform numbers, we can't recover the "true average."
- Bandwagon effect:Earlier reviews influence later ones, so individual ratings aren't necessarily independent.
- Low-sample places:Some places have only a few dozen reviews total, so post-hoc calibration is rougher and directional at best; when the sample is too small we label it "for reference only."
Quoted reviews and ratings are from third-party sources; copyright belongs to the original authors or platforms. This service compiles publicly available information for reference only. See Disclaimer.